Predicting Sales - Forecasting

Restaurant Site Selection

Load packages and R Utility programs:

Loading data and examining data structure.

'data.frame':   33 obs. of  4 variables:
 $ sales      : int  107919 118866 98579 122015 152827 91259 123550 160931 98496 108052 ...
 $ competition: int  3 5 7 2 3 5 8 2 6 2 ...
 $ population : int  65044 101376 124989 55249 73775 48484 138809 50244 104300 37852 ...
 $ income     : int  13240 22554 16916 20967 19576 15039 21857 26435 24024 14987 ...
NULL

Printing data:

Let’s examine the 5 summary statistic:

     sales         competition      population         income     
 Min.   : 91259   Min.   :2.000   Min.   : 37852   Min.   :13240  
 1st Qu.:105564   1st Qu.:3.000   1st Qu.: 57386   1st Qu.:16839  
 Median :122015   Median :4.000   Median : 95120   Median :19200  
 Mean   :125635   Mean   :4.394   Mean   :103887   Mean   :20553  
 3rd Qu.:140791   3rd Qu.:6.000   3rd Qu.:139900   3rd Qu.:22554  
 Max.   :166755   Max.   :9.000   Max.   :233844   Max.   :33242  

Let’s create a histogram data visualization of Sales Frequency:

Plotting Proportion of Restaurants with Lower Sales:

Plotting Correlation Matrix:

Plotting correlation heat map:

Our Regression Model (3 IV’s and Sales as our response variable):

Fitting our linear regression model:

Examining the fitted summary:

Call:
lm(formula = restdata_model, data = restdata)

Residuals:
   Min     1Q Median     3Q    Max 
-21923  -8627  -2956   5328  33887 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept)  1.022e+05  1.280e+04   7.984 8.35e-09 ***
competition -9.075e+03  2.053e+03  -4.421 0.000126 ***
population   3.547e-01  7.268e-02   4.880 3.54e-05 ***
income       1.288e+00  5.433e-01   2.371 0.024623 *  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 14540 on 29 degrees of freedom
Multiple R-squared:  0.6182,    Adjusted R-squared:  0.5787 
F-statistic: 15.65 on 3 and 29 DF,  p-value: 3.058e-06

Examine multicollinearity across explanatory variables to ensure all values are low (say, < 4)

competition  population      income 
   2.348446    2.496186    1.180772 

Plotting Residuals:

Defining the data frame of sites for new restaurants

Next, obtain predicted sales for the new restaurants rounding to the nearest dollar.